Dating of gravelly fluvial deposits is typically hampered by the absence of suitable plant fragments for radiocarbon dating or sand layers for luminescence dating. Although gravelly fluvial deposits generally contain sand-sized quartz and feldspar that could be used for luminescence dating, the considerable grain-to-grain variability in beta dose rate, coupled with the lack of methods for routine dose rate estimation, presents challenges. In this study, a recently proposed model for calculating average beta dose rates in granular matrices was modified to be applicable to sand-sized dosimeter grains within gravelly deposits. We applied the modified model to sand-sized K-rich feldspars within sand matrices obtained from gravelly fluvial deposits in the Tokachi Plain, northern Japan, and compared the ages of samples obtained from sand matrices with those from sand lenses. Although the weight of <2 mm grains accounted for only 20%–35% of the bulk sediment, these grains were estimated to contribute approximately 70% of the external beta dose rate according to the model, because larger grains have a larger self-dose. Taking into account that the beta dose to dosimeter grains is mainly derived from smaller matrices (e.g. <2 mm), beta dose rates were also calculated based on the infinite matrix dose rate of the <2 mm fraction, along with the conventional water correction method. The ages of sand matrices calculated based on the beta dose rates derived from both the model and the infinite matrix dose rate of the <2 mm fraction were generally consistent with those of sand lenses. The dose rate calculated based on the model might be more accurate than that calculated using the infinite matrix dose rate of <2 mm fraction, but calculating the beta dose rate using the infinite matrix dose rate of <2 mm fraction is useful as a simple approach.
Read full abstract